News for hackers. Stuff that matters.

A comment on The Era of 1-bit LLMs: ternary parameters for cost-effective computing. Parent comment

Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by sigmoid10 on Wednesday February 28, 2024 @08:03AM (#39537308)
People are working on that [1]. In some sense, it's a step back to analog computing. Add/multiply is possible to do directly in memory with voltages, but it's less versatile (and stable) than digital computing. So you can't do all calculations in a neural network that way, meaning some digital components will always be necessary. But I'm pretty sure analog will make a comeback for AI chips sooner or later.

[1] https://www.nature.com/articles/s41586-023-06337-5

[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by zcw100 on Wednesday February 28, 2024 @09:15AM (#39538199)
Reminds me of my father saying something about how vacuum tubes are great integrators.
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by monocasa on Wednesday February 28, 2024 @10:48AM (#39539490)
Chips are too. Opamps can add, multiply, subtract, divide, integrate and differentiate depending on how they're plugged in.
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by klysm on Wednesday February 28, 2024 @01:46PM (#39542062)
Hence the name 'operational' amplifier
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by irrelative on Wednesday February 28, 2024 @11:34AM (#39540290)
Hadn't thought about it this way before, but given that LLMs are auto regressive (use their own data for next data), they're sensitive to error drift in ways that are rather similar to analog computers.
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by trebligdivad on Wednesday February 28, 2024 @12:51PM (#39541383)
Trinary however is an interesting middle; people have built trinary hardware long ago; it feels like you could make natively trinary hardware for something like this; it might even be quite a win.
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by int_19h on Wednesday February 28, 2024 @03:02PM (#39542784)
People haven't built reliable ternary electronics, though. Soviets tried with Setun, but they eventually had to resort to emulating each trit with two hardware bits (and wasting one state out of the possible four).
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by eru on Wednesday February 28, 2024 @09:11PM (#39545833)
If you are are using two bits anyway, you might as well represent (-2, -1, 0, 1) instead of ternary?
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by int_19h on Wednesday February 28, 2024 @10:31PM (#39546144)
Sure, but then you lose the symmetry that makes trits so convenient for many things.
[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by thsksbd on Wednesday February 28, 2024 @01:24PM (#39541814)
Can you make a "CMOS" three voltage level circuit though? One where the only current flow is when the state changes?

Im not in this field but that's a question that's been bugging me for a while. Off you can't do this wouldn't energy consumption balloon?

[ Parent | on HN ]
Re:The Era of 1-bit LLMs: ternary parameters for cost-effect... by neomantra on Thursday February 29, 2024 @07:41AM (#39548582)
My friend was working on this in the mid-90s at Texas Instruments. Not sure what the underlying semiconductors were, but it did involve making ternary logic via voltage levels. Just searched a bit and found this TI datasheet which might be an example of it (high logic, low logic, high impedance), but maybe not: https://www.ti.com/lit/ds/symlink/sn74act534.pdf
[ Parent | on HN ]